Absorption Efficiency of Spiral Gas-Lift Wash Bottle B. B. CORSON, Universal Oil Products Company, Riverside, 111.
HE wash bottle shown in
--p 1 ~
' ,
1
A
FIGURE 1
Figure 1,which has been found t o be a useful laboratory tool, is similar to absorbers described by Weaver and Edwards ( 7 ) , Keller (2), Milligan (5), Friedrichs ( I ) , Martin and Green (I), Shaw (6),andMartin (3). It gives intimate and prolonged contact between gas and liquid with recirculation and economy of liquid reagent. The ratio of gas to liquid in the mixture being pumped through the spiral can be yaried by changing the liquid This bottle is especially suitable for scrubbing gas with the 2-phase liquid sysf bromine and water. The gas pumps alternate slugs of bromine and water up the spiral and out the top,
where the bromine falls through the water to the bottoni of the bottle and is immediately recirculated. Consequently, not only is the gas thoroughly scrubbed b y the bromine water but the water is continuously in contact with bromine. The Friedrich bottle, on the other hand, allows the bromine layer to remain more or less stagnant a t the bottom of the bottle, and consequently the water is not continuously maintained saturated with respect to bromine. The bottle can be made in different sizes, but a certain relationship must hold between lengths S and 5" to keep t'he gas from short-circuiting through holes U rather than passing up spiral V . In the usual bottle, S is 60 mm., T is 80 mm., and the spiral consists of 11 turns of tubing 6 mm. in outside diameter (3.5 mm. in inside diameter). The volume capacity of the spiral is 11 cc. The gas delivery tube is constricted at W to about 3-mm. inside diameter. The outer shell of the bottle is about 40 mm. in outside diameter, Alternate slugs of gas and liquid ascend the spiral and are in contact over a length of 950 mm. The back pressure, lioiTever, is no greater than that of an ordinary bottle of the Drechsel type when the length of contact beheen gas and liquid is only about 50 mm.
Experiment a1 At first i t was thought that the contacting efficiencies could be compared on the basis of the amount of R-ater removed by a stream of gas. However, this method RW not applicable because of the disturbing influence of entrainment.
NOVEMBER 15. 1938
ANAL1 TIC 41, EDITIOS
With the spiral bottle, the amounts of water removed per liter of nitrogen a t the gas rates of 0.6, 1.4, 2.8, 5.6, 9.9, and 15.3liters per hour were 0.01~7,0.0187,0.0180,0.0173,0.0173, and 0.0173 gram, respectively. K i t h the Drechsel bottle, the amounts removed at the same gas rates were 0.0177, 0.0172, 0.0169, 0.0175, 0.0184, and 0.0185 gram, respectively. The saturator temperature was 20" C., the pressure within the saturator was 754 nim., and the volume of the nitrogen was converted to 20' and 760 mm. The theoretical amount of water was 0.0178 gram. TABLE I. ABSORPTION EFFICIENCIES .4T COSSTAKT GAS RATE .4SD DIFFEREXT CAUSTIC CONCENTRATIONS Weight
K??H
Total Cog Passed Dreohsel Spiral Grams Grams 1.096 1.200 1.119 1.132 1.119
1; D
4 3 2
1.119 1.114 1.105 1.115 1.119
Unabsorbed C o r Drechsel Spiral Gram Gram 0.0419 0.1161 0.1399 0.2042 0 4263
0.0010 0.0018 0.0018 0,0143 0.2093
E5ciency Drechsel Spiral
%
%
96.2 90.3 87.5 82.0 61.9
99.9 99.9 99.: 98.i 81.3
647
Improved Vacuum Regulator CLAYTON W. FERRY Burroughs Wellcome & Co., Tuckahoe, X. Y.
0
F THE several devices designed to maintain a con-
stant pressure for vacuum distillation, that described by Ellis (1) has proved to be the most satisfactory. However, at pressures of 3 mm. or less, even the small variations in pressure that this regulator allows cause appreciable changes in the boiling point of materials being distilled, and some distillations tend t o build up pressures in the portion of the system beyond the capillary. This capillary, while equalizing and smoothing out variations in pressure caused by the intermittent operation of the pump, also slows up the evacuation of the distillation system, occasionally causing a lag after the distillation is started.
The following method was finally adopted: A mixture of about
99 per cent of nitrogen and 1 per cent' of carbon dioxide was passed through 100 cc. of caustic in the bottle being tested and the weight of carbon dioxide escaping absorption was determined. The setup is shown in Figure 2. The gas in cylinder A was passed at constant rate through bubble-counter B into the caustic in bottle C. The exit gas from C was dried by calcium chloride D
and by Anhydrone E, and the carbon dioxide was absorbed by Ascarite F and G. The water vapor from the Ascarite was retained by Anhydrone H , whence the gas passed through protector tube I containing Anhydrone, Ascarite, and calcium chloride, and finally through meter J. The absorption tubes were weighed against a counterpoise with the usual precautions. Anhydrone H was necessitated by the vapor pressure of Ascarite. The average weight increase of Anhydrone H w&s 0.0098 gram, corresponding to a vapor pressure for Ascarite of 0.16 mm., which is significant when large volumes of gas are involved.
Table I present's data on the relative efficiencies of the two bottles, using different concentrations of caustic. The nitrogen-carbon dioxide mixture contained 0.01736 gram of carbon dioxide per liter of nitrogen (20" C., 760 mm.). The gas rate was 3.8 liters per hour and approximately the same volume of gas was passed in each experiment (63.5 liters of exit nitrogen; 20' C., 760 mm.). Volume of caustic was 100 cc. in each run. TABLE11. ABSORPTIOSEFFICIESCIESWITH 4 PER CEKT PoTASSIUY HYDROXIDE AT DIFFEREST Gas RATES Gas Rate Liters/hr. 3.8 7.6 11.5
Total COz Passed Drechsel Spiral Grams Grams 1.119 1.228 1.228
1.105 1.243 1.226
Unabsorbed COz Dreohsel Spiral Gram Gram 0.1399 0.1830 0.2033
0.0015 0.0050 0.0072
Efficiency Drechsel Spiral
%
%
87.5 85.1 83.4
99.9 99.6 99.4
Table I1 presents data on absorption efficiencies at the same concentration of caustic b u t with different gas rates. The nitrogen-carbon dioxide mixture used a t the gas rates of 7.6 and 11.5 liters per hour contained 0.01943 gram of carbon dioxide per liter of nitrogen (20" C., 760 mm.).
-4cknowledgmen t The authors express thanks t o F. L. Hayes for the glass blowing and to William Cerveny for analytical work.
Literature Cited (1) Friedrichs, Chem. Fabrik, 4, 203 (1931). (2) Keller, Chem-Ztg., 47, 506 (1923). (3) Martin, IKD. Eh-c;. C H m f . ,Anal Ed., 8, 395 (1936). (4) Martin and Green, Ibid.,5, 114 (1933). (5) Milligan, Sczence, 63, 363 (1926). (6) Shaw, ISD. EXG.CHEM.,Anal. Ed., 6, 479 (1934). (7) Weaver and Edwards, J. ISD. ESG.C m v . , 7, 534 (1915) ; B , erin-
stof-Chem., 18, N 9 1 (1937). RECEIVED Tu15 24, 1938
The first of these objections can be overcome by tilting the arm of the regulator, in which contact between the mercury and the electrode is made, from vertical to almost horizontal, as is shorn in the accompanying sketch. This causes the meniscus to travel a greater distance per unit change in pressure. Tubing 4 to 5 mm. in inside diameter is small enough to decrease any tendency of the mercury column to oscillate. If larger tubing is used, i t is advisable to p u t a constriction in the bottom of the U-shaped portion. The mercury terminals for the mires leading t o the relay are designated by a and b, xhile c leads to the pump and d is a pivot for the mounting board. As mercury tends to stick to the platinum or tungsten wires generally used for these contact electrodes, thus reducing the sensitivity, i t n-as necessary to use Chromel wire for the actual contact. T o facilitate the making of a good glass seal, the Chromel nire was soldered to a piece of tungsten nire, which v-as in turn sealed into the side arm. d condenser across the electrodes of the regulator reduces sparking and prevents fouling of the Chromel contact. To operate, evacuate with the stopcock open until approximately the desired pressure is reached, close the stopcock and make the fine adjustment by tilting the assembly board on it; pirot 9s desciihed by Ellis. The hasp of a Bunsen